<p>In response to the current problems such as the shortage of experimental resources, high safety risks in the experimental environment, and insufficient cultivation of students' engineering practice abilities in relevant courses of hoisting and conveying equipment in higher education, a digital twin experiment platform is designed and implemented for crane bridge structures to address the dual needs of teaching and research in this paper. The designed digital twin experiment platform is based on a five dimensional system framework: integrating application modules such as physical experiments, sensor networks, data acquisition and edge processing, virtual simulation, and 3D visualization. Special tasks include the following aspects: the bridge physical and virtual geometric models based on 3D modeling and 3D printing technologies are constructed, high-fidelity visual twin bodies in Unity 3D are realized, and the LabVIEW and NI data acquisition hardware are used to achieve real-time measurement and pre-processing of multi-channel signals such as acceleration and strain, then a two-way communication and synchronous control mechanism between a programmable logic controller, human–machine interaction interface, virtual twin bodies and the data layer is built through industrial protocols such as S7.net and ModBus TCP (TCP, Transmission Control Protocol), and the functions such as data storage, signal analysis, state monitoring and playback control in the service layer and application layer are realized. The designed digital twin experiment platform has realized basic experiments on the kinematics, vibration characteristics and structural health monitoring of bridge structures of the crane, and it is significantly superior to traditional physical experimental methods in terms of safety, cost and repeatability. The article also proposes several technical paths for scientific research (These technological paths are also the research directions that need to be further explored in the future of this study), that including real-time coupling based on the finite element to reduced-order model (FE, finite element→ROM, reduced-order model), data assimilation using UKF(UKF, Kalman Filter)/particle filtering to improve the robustness of online state estimation, and data-driven fault prediction and remaining life estimation based on LSTM(LSTM, Long Short-Term Memory)/Transformer to enhance the platform's prediction and decision-making capabilities. This innovative approach has opened up a new path for experimental teaching and related scientific research practices in the field of crane structures in mechanical engineering, aligning experimental teaching with the development of related scientific research and frontal subject, and it should be to promote engineering education to keep pace with the times.</p>